The present invention relates to a technique for drive control of a brushless motor and a technique effective when applied to a method for determining phases (a pair of phases) at which to start current conduction when starting the motor, and particularly concerns a technique effective when used in a drive control apparatus of a spindle motor for rotating a disk-type storage medium, such as a HDD (hard disk drive) device.
With hard disk devices, there has been a strong demand for higher speed in writing and reading information on a magnetic disk, namely, quicker access speed. To this end, it is required that the spindle motor be made much faster. In addition, demand is also mounting for reductions in size, power consumption and production cost of the drive control apparatuses. In conventional hard disk devices, DC polyphase brushless motors are generally used for their spindle motor to rotate the magnetic disks at high speed, and information is written or read on the rotating magnetic disk by bringing the read/write magnetic heads into contact with or in close vicinity to the disk.
In brushlress motors, there has been used a motor drive control method by which to prevent reverse rotation of the motor by detecting the positional relation of the rotor and the stator by means of Hall elements and by, from the detected positional relation, determining field-coil phases at which current conduction is to be started. Because mounting a rotor position detector using Hall elements in the motor increases the difficulty of downsizing the motor, sensorless motors have come to be used in large numbers in the hard disk devices. However, if the magnetic disk is driven by a sensorless motor, the rotor is likely to make a reverse rotation for an instant with a probability of {fraction (1/2+L )} when the disk starts to rotate. 
With the rapidly multiplying storage density of the magnetic disks in hard disk devices in recent years, the magnetic read/write heads have been sharply reduced in size. Consequently, in the hard disk devices with the magnetic heads miniaturized to such an extent, there is a problem that if the rotor is turned in reverse even for an instant, the magnetic heads may suffer a fatal damage. To solve this problem, a control method has been proposed in which a pulse current of so short a duration as not to cause the rotor to react is supplied to the field coils of the stator, and the field coils where the amplitude is at the maximum value, in other words, the phases, where the field of the rotor magnet in the same direction as the generated field of the coils, causing magnetization to be saturated to make current flow most easily, are determined as the phases at which to start current conduction (Refer to JP-A-63-694895 published on Mar. 29, 1988 which corresponds to U.S. Ser. No. 880754 filed on Jul. 1, 1986).
Another control method has been proposed in which a pulse current is conducted through the field coils of the stator and then the pulse current is conducted in the opposite direction, and differences in current rise time constant are detected at respective field coils where the current is passed through, and according to detection results, the position of the rotor is determined to determine a pair of phases at which current conduction is started. In other words, this control method is such that phases at which current conduction is started are determined by determining the rotor position based on detection results obtained by detection of differences in inductance by making use of a phenomenon that the inductance of the field coils varies whether the direction of the magnetic field is the same or not between the field coils and the rotor magnet (that is to say, whether magnetic saturation occurs or not) (Refer to JP-A-3-207250 published on Sep. 10, 1991 which corresponds to U.S. Ser. No. 413311 filed on Sep. 27, 1989).
In addition to the above inventions, another invention has been proposed that the stopped position of the rotor is determined by applying a diagnosis signal of a frequency higher than the frequency of an exciting signal applied when the motor is started, to a single coil or two or more coils connected in series and detecting an induced voltage of one of the serially-connected coils (Refer to JP-A-7-274585 published on Oct. 20, 1995).
However, the present inventors have revealed that the prior art described above suffer problems as at follows.
In the control method that determines a pair of phases, where current conduction is started, by passing a pulse current and detecting the maximum amplitude value, the maximum amplitude value depends on variations in winding in the field coils of the stator, for which reason detection errors occur due to very small winding variations that are unavoidable in the manufacturing process. In the control method that determines a pair of phases, where current conduction is started, by detecting the rotor position based on differences in current rise time constant, because a phenomenon of magnetic saturation is used, differences in time constant do not become conspicuous unless a fairly large current is passed, and therefore it is difficult to detect differences in the time constant when a current passed is so small as the rotor does not react to it. Another problem with this control method is that the point of reversal of the large-small relation among the time constants that occurs when the direction of a current is reversed does not coincide with the point of magnetic saturation, resulting in errors in determination results.
The present invention has as its object to provide a brushless motor drive control technique that can prevents reverse rotation of the motor at starting by detecting the position of the rotor relative to the stator with fewer errors and determining a field coil pair at which current conduction is started.
According to an aspect of the present invention, a pair of phases for current conduction to start the motor is determined by passing a pulse current with a duration so short as the rotor does not react through the field coil of any phase of the motor in first and second, mutually opposite, directions sequentially, and detecting induced voltages in the non-conducting phase by a pulse current in two opposite directions, combining voltages induced by a pulse current in the first direction and a pulse current in the second direction, detecting the polarities of combination results, and determining a pair of phases for current conduction when starting the motor based on polarity detection results related to a plurality of the conducting phases.
The above-mentioned and other objects and features of the present invention will become obvious from the following description of this specification and the accompanying drawings.